TraceGraph AI — Autonomous Web Crawler, Knowledge Graph & PR Blast Radius

Autonomous Crawler & Exploration Center

Agentic browser exploration: observes live screens, extracts interactive candidate actions, computes state fingerprints, and builds concrete transitions.

Target Application & Parameters

Target Storefront / Web App URL

Custom public URL

Link Depth: 4

Max Actions: 40

Max Screens: 20

Capture Screenshots Capture DOM Autonomous Exploration

Live crawl only. Up to 25 distinct same-domain screens; records full-page browser artifacts and does not substitute fixture data.

IDLE / READY

00:00

0 Pages

0 States

0 Actions

0 Transitions

CURRENT SCREEN: —

CURRENT ACTION: —

Live Crawler Event Stream

Ready to initiate autonomous exploration session...

Discovered Screen Artifacts & Screenshots

No crawl selected

Transition ID	From Screen	Trigger Action	To Screen	Interaction Type

Ingest Product Requirements

Choose one source: a public README/documentation URL or pasted requirement text. The system extracts testable product requirements.

Choose a Requirement Source

Use one option only. A URL is fetched from an approved public documentation host; pasted text stays in this request.

Public GitHub README / documentation URL Paste Markdown / requirement text

Public documentation URL Loading approved documentation hosts…

Ingestion Summary & Source Provenance

Requirements Extracted: No source selected

Provenance Mode: Awaiting source

Inferred category: —

Average Testability Score: —

Structured Requirements Table

Req ID	Category	Requirement Text	Testability	Coverage Status

Past Autonomous Crawl Sessions

Audited record of all browser explorations, states captured, and screen graphs.

Crawl ID	Target URL	Status	Discovered Pages	Transitions	Started At	Actions

PR Code Symbol UI Element Page Flow Requirement

Deterministic PR Impact Breakdown

Derived from graph traversal paths with hop-weight confidence arithmetic.

Live AI Executive Blast Radius Summary

EVIDENCE REQUIRED

No provenance-verified report is loaded. Complete the crawl, specification ingestion, and graph build steps for the selected repository and PR.

Impacted UI Components

Element Label	Type	Risk Tier	Confidence	Evidence Path

Affected User Flows

Flow ID	Flow Name	Risk Tier	Confidence

Requirements at Risk

Req ID	Requirement Text	Category	Confidence	Coverage Status

Evidence-Grounded QA Plan

Tests are generated only from a provenance-verified PR report, a completed crawl, captured DOM/screenshots, and browser-observed transitions.

QA Analysis Input

No QA analysis loaded. Select a completed crawl after a provenance-verified PR analysis.

Agent Execution Trace

Agent	Status	Evidence Output

Evidence Verification

Claim	Status	Confidence	Reason

Reviewer-Ready Test Cases

ID	Test	Priority	Review Status	Evidence

What We Do Not Know

Mathematical Evidence Paths (No Hallucinations)

Every single risk claim is proven via concrete graph hops from PR diff → Code AST → UI selector → User journey.

4-State Requirement Coverage Matrix

Models what should exist versus what is observed in the UI (including explicit absence modeling).

COVERED: 0 PARTIAL: 0 UNVERIFIED: 0 ABSENT: 0

Req ID	Category	Requirement Description	Testability	Coverage Status	UI Evidence